Japanese localization | Tech Off

Larsenal
ready to give an answer

I'm localizing a site into Japanese( and French).

First, what's a good text editor for this sort of thing? I've been using EditPlus for years, but it renders the Japanese text as the ugly boxes.

Second, having never done a Japanese site, is there anything tricky to getting a page to display Japanese correctly?

Duncanma
Just Coding for Fun...

I'm assuming (because of the editor question) that you are creating copies of the actual web pages in these different languages, not localizing strings in a database or in a resource file or something like that...

Honestly, Notepad will do it.. but only if you are careful to always save as unicode... for some reason, it seems to default to ASCII ... and that just really sucks. I'm sure something like Visual Studio Express (web) would handle it as well...

Larsenal
ready to give an answer

In reply to Duncanma

#Jun 20th, 2008 @ 4:52 PM

Hmn... tried notepad, but when I paste from Excel (which is where they have the side-by-side copy) I still get the blocks. Made sure it was being saved as Unicode, but no luck.

Tried VS2005, and it shows up fine. This will be a rare opportunity to edit PHP in VS2005.

Thanks, Duncan!

Sven Groot
My name has 9 letters. Coincidence? I think not...

In reply to Larsenal

#Jun 20th, 2008 @ 5:06 PM

Be careful when using Notepad for unicode (both utf-8 and utf-16). It adds a byte-order mark to the start of the file, and not all web browsers like that.

CannotResolveSymbol
{insert caption here}

In reply to Larsenal

#Jun 20th, 2008 @ 5:06 PM

Also of interest may be Expression Web 2: it supports Unicode correctly and it has full-blown support for PHP (with proper highlighting and intellisense!).

Larsenal
ready to give an answer

In reply to CannotResolveSymbol

#Jun 21st, 2008 @ 12:52 PM

What happened to my CSS file?

So I throw up a simple little HTML page with some Japanese copy... saved it as Unicode and uploaded to server. Strangely enough, FF3 now mangles the external CSS file. The page renders as if no styles are applied. Firebug's Net tab shows the contents of the CSS file... and it's all in (presumably) Japanese. IE handles it fine. I don't know whether this another case of Microsoft's graceful "non-compliance" or whether I've mucked something up.

It's likely that the CSS files (which someone else did) and JS are all ANSI. Is it incorrect to mix ANSI and Unicode (in different files, of course) for a single page rendering?

Sven Groot
My name has 9 letters. Coincidence? I think not...

In reply to Larsenal

#Jun 23rd, 2008 @ 1:31 PM

That shouldn't be a problem. Could you check (with Fiddler or something) if the server is indicating a charset for the CSS file?

Larsenal
ready to give an answer

In reply to Sven Groot

#Jun 23rd, 2008 @ 9:04 PM

Very odd. I didn't see any charset anomolies in the raw HTTP traffic.

Check out my test page in FF. I've only tested in FF3 so far, but I'd venture to guess this isn't a new behavior.

CannotResolveSymbol
{insert caption here}

In reply to Larsenal

#Jun 24th, 2008 @ 10:29 AM

edit: Channel 9 is eating my code!!!

Your server isn't sending a character set for either file, so Firefox has to guess the character set. By default, Firefox uses the same character set as the page that's serving it (miraculously, Firefox is detecting this correctly Wink

).

You can fix this issue in three ways:
(1) Make your web server send the correct character set in the HTTP headers. Read the Apache docs on how to do this.

(2) Specify the character set in the <link> tag. See http://www.w3.org/TR/1999/REC-html401-19991224/struct/links.html#edef-LINK

(3) Author your document using UTF-8 instead of Unicode-16. UTF-8 is backwards compatible with the ASCII character set (meaning you don't have to do either of the above) and can represent all characters in the Unicode character set. In my opinion, this is the "right" way to do it.

Larsenal
ready to give an answer

In reply to CannotResolveSymbol

#Jun 24th, 2008 @ 10:50 AM

Thanks for all the help. Problem solved.

(Yesterday, I only skimmed this helpful post, but should have read it thorougly as it would have cleared up the issue.)

Sven Groot
My name has 9 letters. Coincidence? I think not...

In reply to CannotResolveSymbol

#Jun 24th, 2008 @ 10:50 AM

I find my life is a lot easier since I just started saving all my files as utf-8. No worries that way, it always works.